A Group Genetic Algorithm for Energy-Efficient Resource Allocation in Container-Based Clouds with Heterogeneous Physical Machines

Containers are quickly gaining popularity in cloud computing environments due to their scalable and lightweight characteristics. However, the problem of Resource Allocation in Container-based clouds (RAC) is much more challenging than the Virtual Machines (VMs)-based clouds because RAC includes two levels of allocation problems: allocating containers to VMs and allocating VMs to Physical Machine (PMs). In this paper, we proposed a novel Group Genetic Algorithm (GGA) with energy-aware crossover, Best-Fit-Decreasing Insert (BFDI), and Local Search based Unpack (LSU) operator to solve RAC problems. Meanwhile, we apply an energy model with heterogeneous PMs that accurately captures the energy consumption of cloud data centers. Compared to state-of-the-art methods, experiments show that our method can significantly reduce the energy consumption on a wide range of test datasets.